This dataset comprises of spatial boundaries delineating 10956 major retail agglomerations across the United States, with an accompanying classification that describes their characteristics. The dataset was generated through the use of retailer location data supplied by SafeGraph. The data provides a replicable data product built on a heuristic categorisation of retail unit density. The product is built using consistent methods and data for the national extent of the U.S., representing the first delineation of retail centres for this country.

The agglomerations are identified based on the clustering and connectivity patterns of individual retail units over space. A hexagonal high-resolution grid is superimposed over spatial clusters of retail points and a network-based algorithm is used to prune and fine-tune clusters into self-contained, mutually exclusive tracts.

The retail boundaries are accompanied by information about their geographical location, including state, county, place and street names, as well as a non-hierarchical classification which describes the characteristics of the different retail centres. A two-tier classification is presented, comprising four top-level groups and fourteen nested types.

SafeGraph have kindly allowed this CDRC-created dataset, aggregated and derived from their own data along with OpenStreetMap's, to be distributed under an open licence. SafeGraph, OpenStreetMap and CDRC should be attributed when using this data.